土壤水分作为地表系统研究中的关键参数之一，在水循环、农业生产、气候变化以及环境监测研究中是必不可少的参量。针对目前土壤水分光学遥感反演方法中无法直接获取定量土壤水分、在不同大气条件下得到的土壤水分指数无法进行像元间直接比较、像元级根区土壤水分难以获取等关键问题，本项目立足于定量土壤水分遥感反演的机理研究，运用土壤-植被-大气传输模型Noah模拟土壤水分在不同大气和下垫面条件下的垂直分层变化情况，揭示土壤水分时空变化机制与规律；基于地表能量平衡模型和热传导方程，通过理论推导、模型模拟、数学求解、比较验证等手段，发展在不同大气条件下可直接进行像元间比较的土壤水分定量反演模型；结合陆面过程模型数据同化技术与水热传导机理，利用集合卡尔曼滤波理论探求遥感反演表层土壤水分与根区土壤水分的物理定量关系，实现大面积区域根区土壤水分光学遥感估算，为农业灌溉管理、区域干旱监测及预警提供技术服务和决策支持。

As one of the key parameters in surface systems research, soil water content plays an essential role in the water cycle, agriculture, climate change and environmental monitoring studies. Due to the current optical remote sensing technology can not obtain direct quantitative soil moisture and the soil moisture index obtained under different atmospheric conditions cannot be directly compared between pixels as well as the difficulty in estimating soil moisture in the root zone at the pixel level, this study aims to the mechanism of quantitative soil moisture remote sensing inversion and used the soil-vegetation-atmosphere transmission model - Noah with a robust physical basis to simulate the soil moisture variation for vertical layers under different atmospheric and underlying conditions, which can reveal the mechanism and law of soil moisture spatial and temporal change. Based on surface energy balance model and heat conduction equation, a mathematical model between the components of surface flux and remotely sensed parameters was established such as surface temperature, air temperature, vegetation coverage, soil moisture and other variables through theoretical derivation, model simulation, mathematical solving and comparison validation. A new type of normalized feature space was conducted with the theory of surface temperature-vegetation coverage feature space that can remove the influence of atmospheric conditions. Therefore, a quantitative soil moisture inversion model was developed that can be directly compared between pixels on different days. Combined with data assimilation technology of land surface process model and the water and heat transfer mechanism, the ensemble Kalman filter theory is used to explore the physical quantitative relationship between remotely sensed surface soil moisture and root zone soil water content to achieve the soil moisture estimation in the root zone of large areas with optical remote sensing data, which can provide technical services and decision support for agricultural water resources management and regional drought monitoring.